Device state monitoring system and state monitoring method
By combining device monitoring with SNS functions, the system automates the analysis and sharing of equipment data, reducing the burden on administrators and enabling proactive detection of abnormalities, thus enhancing operational efficiency and reducing downtime.
Patent Information
- Application Number
- JP2023081658
- Authority / Receiving Office
- JP · JP
- Patent Type
- Applications
- Current Assignee / Owner
- Filing Date
- 2023-05-17
- Publication Date
- 2024-11-28
AI Technical Summary
The burden of monitoring and analyzing vast amounts of data from mechanical equipment operation logs for abnormalities and sharing maintenance information among users is significant, making it difficult to detect potential issues before they become critical.
A system that integrates device condition monitoring with social networking services (SNS) to reduce the burden by using trained models to process and share information, including character strings indicating device conditions and maintenance actions, allowing for automated detection and notification of abnormalities.
This integration reduces the workload of monitoring and analysis, enables easier information sharing, and allows for proactive detection of abnormalities, minimizing downtime by notifying administrators and facilitating timely maintenance actions.
Smart Images

Figure 2024165446000001_ABST
Abstract
Description
[Technical field]
[0001] The present invention relates to a device status monitoring system and a status monitoring method. [Background technology]
[0002] Conventionally, various mechanical devices such as bearing devices and sliding devices are monitored for their status, and the mechanical devices are controlled according to the detected status to continue to operate appropriately. When a status change is detected, a configuration is known in which a user such as an administrator is notified according to the detected status, and maintenance or adjustment of operating conditions is encouraged as necessary. For example, Patent Document 1 discloses a configuration in which a status change of a device is monitored, and an e-mail is sent to a specified address according to the detected status.
[0003] In addition, in the control of a mechanical device, a configuration is also known in which instructions are given using predetermined control parameters, as well as natural language. For example, Patent Document 2 discloses a configuration in which, in order to improve user convenience, requests including natural language are accepted as operations for the device, and keywords are extracted from the requests, and then control of the device is executed. [Prior art documents] [Patent documents]
[0004] [Patent Document 1] JP 2002-215484 A [Patent Document 2] JP 2014-104625 A Summary of the Invention [Problem to be solved by the invention]
[0005] In recent years, social network services (SNS) have become popular as a tool for various users to share information. For example, a maintenance worker who has performed maintenance on a device can register the details of the maintenance and information on the target device in the SNS service, and other users can access the SNS service to refer to this information. SNS services have various functions, and for example, it is possible to register any message in natural language, etc.
[0006] On the other hand, when monitoring the status of the above-mentioned devices based on the operation logs of the machines, the amount of data becomes huge, so it is a heavy burden for the administrator to analyze each log and determine whether there is an abnormality. Also, it is assumed that the operation logs before maintenance of the machines, that is, immediately before an abnormality is detected, are useful for detecting an abnormality or a sign of an abnormality, but it is also a heavy burden for the administrator to monitor and analyze them each time.
[0007] In view of the above problems, the present invention aims to provide a system that combines functions related to device status monitoring with functions based on SNS, thereby reducing the burden of device status monitoring, information sharing between users, and information analysis. [Means for solving the problem]
[0008] In order to solve the above problems, the present invention has the following configuration: That is, a status monitoring system for a monitored device, the status monitoring system comprising: a first system for providing a social networking service; a second system for determining a state of the monitored device using measurement data during operation of the monitored device; The present invention relates to a method for manufacturing a computer-implemented ... The first system comprises: A means for accepting a posting of a first message including first character string information indicating a state occurring in the monitored device and second character string information indicating a maintenance operation performed in response to the state; a means for providing posted messages to other systems; A means for outputting a message posted on the social networking service; having The second system comprises: means for acquiring measurement data during operation of the monitored device; A means for acquiring first character string information and second character string information corresponding to new measurement data using a trained model that receives measurement data as input and outputs first character string information and second character string information corresponding to the measurement data; means for posting a second message to the first system, the second message including the acquired first string information and the second string information; having A condition monitoring system, wherein the trained model is generated by learning using training data composed of first string information and second string information contained in a message acquired from the first system and the measurement data.
[0009] Another aspect of the present invention has the following configuration. A first system for providing a social networking service; a second system for determining a state of the monitored device using measurement data during operation of the monitored device; A status monitoring method for a status monitoring system comprising: In the first system, receiving a first message including first character string information indicating a state occurring in the monitored device and second character string information indicating a maintenance operation performed in response to the state; providing the posted messages to another system; outputting a posted message on the social networking service; having In the second system, acquiring measurement data during operation of the monitored device; A step of acquiring first character string information and second character string information corresponding to new measurement data using a trained model that receives measurement data as input and outputs first character string information and second character string information corresponding to the measurement data; posting a second message to the first system, the second message including the acquired first string information and the second string information; having A condition monitoring method, wherein the trained model is generated by learning using training data composed of first string information and second string information contained in a message acquired from the first system and the measurement data. Effect of the Invention
[0010] The present invention makes it possible to combine functions related to device status monitoring with functions of an SNS, thereby reducing the burden of device status monitoring, information sharing between users, and information analysis. [Brief description of the drawings]
[0011] [Figure 1] 1 is a block diagram showing an example of the configuration of a status monitoring system according to an embodiment of the present invention; [Diagram 2] 1 is a block diagram showing an example of the configuration of an information processing device according to an embodiment of the present invention. [Diagram 3] FIG. 4 is a sequence diagram relating to data generation of the status monitoring system according to the embodiment of the present invention. [Figure 4] FIG. 4 is a sequence diagram relating to a learning process of the state monitoring system according to the embodiment of the present invention. [Diagram 5] FIG. 2 is a sequence diagram relating to a state monitoring process of the state monitoring system according to the embodiment of the present invention. [Figure 6] FIG. 2 is a block diagram for explaining the functional configuration (learning) of the state monitoring system according to one embodiment of the present invention. [Figure 7] 1 is a block diagram for explaining a functional configuration (status monitoring) of a status monitoring system according to an embodiment of the present invention. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
[0012] Hereinafter, an embodiment for carrying out the present invention will be described with reference to the drawings. Note that the embodiment described below is one embodiment for explaining the present invention, and is not intended to be interpreted as limiting the present invention, and all configurations described in each embodiment are not necessarily essential configurations for solving the problems of the present invention. In addition, in each drawing, the same components are given the same reference numbers to indicate their correspondence.
[0013] <First embodiment> The first embodiment of the present invention will be described below. The system according to this embodiment includes a status monitoring system 1, a monitored device 500, a maintenance terminal 600, and a management terminal 700.
[0014] The condition monitoring system 1 includes a data collection system 100, an SNS system 200, a data management system 300, and a learning system 400. Each system constituting the condition monitoring system 1 may be configured as an on-premise or cloud-based server. Therefore, each system constituting the condition monitoring system 1 may be realized by one or more server devices, or may be realized together by one device. The configuration of each system is an example, and other systems may be included, or some systems may be omitted or integrated. When each system is configured as a web-based service such as a cloud-based service, each system may be configured as a separate service and may be configured to be able to cooperate with each other.
[0015] In this embodiment, "status monitoring" includes capturing a sign of a status change, for example, a sign of an abnormality, in addition to a status that has already occurred in a device (here, one or more monitored devices 500). In this embodiment, the "abnormality" is not particularly limited, and may be a request to stop the device itself or a request to switch any control parameter or operating condition in the device. Therefore, these may be changed depending on the configuration of the monitored device 500.
[0016] The data collection system 100 appropriately collects data on the status of each device from one or more monitored devices 500. The content of the data collected here is not particularly limited, and may be changed depending on the configuration of the monitored device 500 and the monitoring content. Data collection may be configured such that the data collection system 100 is the starting point to make a request to the monitored device 500, and in response, the monitored device 500 transmits the corresponding data. Alternatively, the monitored device 500 may be the starting point to transmit data to the data collection system 100 based on a predetermined condition (for example, a predefined time interval).
[0017] Furthermore, depending on the content of the collected data, the data collection system 100 transmits the data to the data management system 300, or posts information related to the data to the SNS system 200. An example of posting to the SNS system 200 will be described later.
[0018] The SNS system 200 is a system that provides a social networking service that allows multiple users (including devices) to communicate with each other. A user can input (post) predetermined information (messages, etc.) to the SNS system 200, allowing other users to refer to or use the information. Functions provided by the SNS system 200 may include publicly known communication functions in addition to the functions described below.
[0019] The data management system 300 manages the data collected by the data collection system 100. Furthermore, the data management system 300 provides the managed data in response to requests from other systems.
[0020] The learning system 400 performs learning processing using data managed by the data management system 300, and generates a trained model to be used during state monitoring. Details of the learning processing and trained model used in this embodiment will be described later.
[0021] The monitored device 500 is a device that is the subject of status monitoring. Although only one monitored device 500 is shown in Fig. 1, a plurality of monitored devices 500 may be provided, and data from each of the monitored devices 500 may be collected by the data collection system 100. The monitored device 500 may be, for example, a processing machine, a polishing device, or the like, and the type and number of the monitored device 500 are not particularly limited.
[0022] The maintenance terminal 600 is an information processing device used by a user (maintenance worker) who performs maintenance work such as maintenance on the monitored device 500. The maintenance terminal 600 may be an information processing device such as a PC (Personal Computer), or may be a mobile terminal such as a smartphone, a tablet terminal, or a POS terminal. Although only one maintenance terminal 600 is shown in FIG. 1, multiple maintenance terminals 600 may be included depending on the number of users who perform maintenance work.
[0023] The management terminal 700 is an information processing device used by a user (device administrator) who manages and monitors the monitoring target device 500. The maintenance terminal 600 may be an information processing device such as a PC (Personal Computer), or may be a mobile terminal such as a smartphone, a tablet terminal, or a POS terminal. Although only one management terminal 700 is shown in FIG. 1, multiple management terminals 700 may be included depending on the number of users who perform management. In addition, the maintenance terminal 600 and the management terminal 700 may be configured as an integrated unit. In addition, the maintenance terminal 600 and the management terminal 700 may be configured as, for example, a general-purpose information processing device such as a PC, or may be configured as a dedicated device.
[0024] In the following description, "learning" or "machine learning" refers to generating a "trained model" by performing learning using training data and an arbitrary learning algorithm. The trained model is appropriately updated as learning progresses using multiple training data, and its output changes even for the same input. Therefore, the trained model is not limited to a state at any point in time. Here, a model used in learning is referred to as a "trained model," and a learning model that has undergone a certain degree of learning is referred to as a "trained model."
[0025] In addition, specific examples of "learning data" will be described later, but the configuration may be adjusted or changed depending on the learning algorithm used. In addition, the learning data may include teacher data used for learning itself, verification data used for verifying a trained model, and test data used for testing a trained model. In the following description, when data related to learning is collectively indicated, it is written as "learning data", and when data when performing learning itself is indicated, it is written as "teacher data". Note that it is not intended to clearly classify the teacher data, verification data, and test data contained in the learning data, and for example, depending on the method of learning, verification, and testing, all of the learning data may also be teacher data.
[0026] In addition, the configuration of the learning data and the preprocessing of the learning data may vary depending on the type of learning algorithm and the contents of the input and output of the learning model. Therefore, the combinations of each data item described below are examples and are not limited to these.
[0027] (Device configuration) FIG. 2 is a block diagram showing an example of the configuration of various systems constituting the status monitoring system 1 according to this embodiment, the maintenance terminal 600, and an information processing device that can be used as the management terminal 700.
[0028] The information processing device 10 includes a control unit 11, a storage unit 12, an operation unit 13, a display unit 14, an external IF (Interface) 15, and a communication unit 16.
[0029] The control unit 11 may be composed of a CPU (Central Processing Unit), an MPU (Micro Processing Unit), a DSP (Digital Single Processor), a GPU (Graphical Processing Unit), a GPGPU (General-Purpose computing on Graphics Processing Units), or a dedicated circuit, etc. The storage unit 12 is composed of volatile and non-volatile storage media such as an HDD (Hard Disk Drive), a ROM (Read Only Memory), and a RAM (Random Access Memory), and is capable of inputting and outputting various information based on a request from the control unit 11.
[0030] The operation unit 13 is a part used when accepting operations from a user, and may be composed of, for example, a mouse, a keyboard, etc. The display unit 14 displays the diagnosis results and an input screen, and is composed of a liquid crystal display, etc. The operation unit 13 and the display unit 14 may be integrated into one device, such as a touch panel display.
[0031] The external IF 15 is an interface for connecting to an external device, and is used, for example, for transmitting and receiving data to and from other systems. The communication unit 16 is a network interface for communicating with an external device, and may use any communication standard, such as wired or wireless.
[0032] The functions of each system described later may be realized, for example, in the information processing device 10 constituting each system, by the control unit 11 reading and executing various programs and data stored in the storage unit 12. Furthermore, the functions of each system described later may be implemented by being installed in the information processing device as an application program, or may be implemented as a Web application that functions through a browser or the like.
[0033] [Processing flow] In this embodiment, the state monitoring of the monitored device 500, which is the monitoring target, and learning using collected data are performed simultaneously in parallel. Note that the trained model used for the state monitoring is generated as a result of a certain learning process, but the learning process may be repeated as appropriate, and the trained model may also be updated. Here, each phase of "data collection", "learning", and "state monitoring" will be described individually.
[0034] (Data Collection) FIG. 3 is a sequence diagram of a process when various data are collected. In this embodiment, measurement data, abnormality information, and maintenance information are used as data to be collected. In this embodiment, the measurement data includes an operation log, a detection value, a control parameter, and the like, which are appropriately recorded in the monitored device 500 during the operation of the monitored device 500. In addition, the measurement data includes time information at the time of measurement. In this embodiment, the abnormality information is a character string to which "# (hashtag)" is added among character string information posted to the SNS system 200 via the maintenance terminal 600 by the user who performed the maintenance work. In the abnormality information, information (classification, etc.) on the abnormality dealt with in the maintenance work by the user who performed the maintenance work is set. In this embodiment, the maintenance information is a character string entered together with the abnormality information among character string information posted to the SNS system 200 via the maintenance terminal 600 by the user who performed the maintenance work. In the maintenance information, information on the content dealt with in the maintenance work by the user who performed the maintenance work (such as the work items performed) is shown as a character string.
[0035] In S301, the monitored device 500 determines whether or not an abnormality has been detected during operation. The content of the abnormality here may differ depending on the function of the monitored device 500, etc. The abnormality to be detected may be predefined on the device side, or may be configurable by a user (e.g., a device administrator). If an abnormality is detected (YES in S301), the processing of the monitored device 500 proceeds to S302. On the other hand, if no abnormality is detected (NO in S301), the processing of the monitored device 500 proceeds to S303.
[0036] In S302, the monitored device 500 transmits data indicating the details of the detected abnormality as measurement data to the data collection system 100. Thereafter, the process of the monitored device 500 returns to S301 and the process is repeated.
[0037] In S303, the monitored device 500 determines whether it is a predetermined transmission timing. The predetermined transmission timing may be a data transmission interval that is specified in advance, or may be when a request is received from the data collection system 100. Alternatively, it may be when a predetermined transmission condition is satisfied. If it is the predetermined transmission timing (YES in S303), the processing of the monitored device 500 proceeds to S304. On the other hand, if it is not the predetermined transmission timing (NO in S303), the processing of the monitored device 500 returns to S301 and the processing is repeated.
[0038] In S304, the monitored device 500 transmits data relating to the operation recorded during the operation as measurement data to the data collection system 100. Thereafter, the process of the monitored device 500 returns to S301 and the process is repeated.
[0039] The processing steps of S301 to S304 are executed in each of one or more monitored devices 500, and are continuously repeated while these monitored devices 500 are operating.
[0040] In S311, the maintenance terminal 600 accepts a work report for the monitored device 500 from the user. The reception here may be performed via a screen that is displayed when the user accesses the SNS system 200 via the maintenance terminal 600 and requests an input screen. The work report here includes the above-mentioned abnormality information (e.g., "#abnormality recovery#motor replacement") and maintenance information (e.g., "XX motor replacement work has been performed"). In addition, the timing of the work report corresponds to, for example, a case where a maintenance worker replaces a part to deal with an abnormality that has occurred in the monitored device 500.
[0041] In S312, the maintenance terminal 600 transmits information based on the report content received from the user by posting it to the SNS system 200.
[0042] In S321, the data collection system 100 stores in a storage unit the measurement data (detection data of anomaly) received from the monitored device 500. The measurement data here includes timestamp information indicating the timing of acquisition.
[0043] In S322, the data collection system 100 posts information related to the abnormality based on the measurement data received in S321 to the SNS system 200. For example, a message (e.g., "#motor overload") including "# (hashtag)" and a character string indicating the abnormality (e.g., "motor overload") is posted to the SNS system 200.
[0044] In S323, the data collection system 100 stores in a storage unit the measurement data (periodic measurement data) received from the monitored device 500. The measurement data here includes timestamp information indicating the timing of acquisition.
[0045] In S324, the data collection system 100 transmits a part or all of the measurement data received in S323 to the data management system 300. The data here may be divided and transmitted according to the amount of data held, or only data necessary for subsequent processing may be extracted and transmitted.
[0046] In S331, the SNS system 200 stores the post received from the data collection system 100. When storing the post, for example, the data format may be converted or predetermined management information may be added in accordance with the specifications within the SNS system 200.
[0047] In S332, the SNS system 200 transmits the data stored in S331 to the data management system 300. The data transmitted here may be, for example, data corresponding to a post regarding anomaly detection, including anomaly information.
[0048] In S333, the SNS system 200 stores the post received from the maintenance terminal 600. When storing the post, for example, the data format may be converted or predetermined management information may be added in accordance with the specifications within the SNS system 200.
[0049] In S334, the SNS system 200 transmits the data stored in S333 to the data management system 300. The data transmitted here may be, for example, data corresponding to a post regarding a work report including anomaly information and maintenance information.
[0050] In S341, the data management system 300 receives data (a post related to anomaly detection) from the SNS system 200.
[0051] In S342, the data management system 300 receives the data from the data collection system 100.
[0052] In S343, the data management system 300 receives data (a post related to a work report) from the SNS system 200.
[0053] In S344, the data management system 300 stores the data received in S341 to S343. The data management system 300 may change the storage format depending on the type, sender, and timing of reception of the received data.
[0054] (study) Fig. 4 is a sequence diagram of a process for performing learning using various collected data and generating a trained model to be used during state monitoring. This process sequence may be performed, for example, when data collection shown in Fig. 3 is performed and a certain amount of data is accumulated as training data. Alternatively, it may be performed at any timing by a system administrator. Since the learning process generally imposes a high load, it is preferable that the process by the learning system 400 is performed separately from the operation of other systems, and the learning system 400 provides the generated trained model in response to a request from the other systems.
[0055] In S401, the data management system 300 transmits data stored in the learning system 400. The data transmitted here may be, for example, only the difference from the previous transmission, or all data within a predefined period may be transmitted.
[0056] In S411 , the learning system 400 receives data from the data management system 300 .
[0057] In S412, the learning system 400 generates first learning data to be used in the first learning process, using the data received in S411 from the data management system 300. Details of the first learning process and the first learning data will be described later with reference to FIG.
[0058] In S413, the learning system 400 generates second learning data to be used in the second learning process, using the data received in S411 from the data management system 300. Details of the second learning process and the second learning data will be described later with reference to FIG.
[0059] In S414, the learning system 400 executes a first learning process for generating a first trained model using the first training data generated in S412. As a result of this process, a trained model for state estimation is generated as the first trained model.
[0060] In S415, the learning system 400 executes a second learning process to generate a second trained model using the second training data generated in S413. As a result of this process, a trained model for generating maintenance information is generated as the second trained model.
[0061] In S416, the learning system 400 saves the first trained model generated in S414 and the second trained model generated in S415. The learning system 400 may also manage data used in the learning, versions of the trained models, and the like.
[0062] (Status monitoring) Fig. 5 is a sequence diagram of a process for monitoring the state of the monitored device 500 using the generated trained model. Before this process is started, each trained model is generated and made available by the sequence in Fig. 4. Note that the sequence shown in Fig. 4 is performed as appropriate, and each trained model is updated as appropriate by the learning process, but the process sequence in Fig. 5 does not particularly limit which stage of the trained model is used.
[0063] In S501, the data management system 300 transmits the measurement data received from the monitored device 500 to the learning system 400. The transmission here may be performed every time the measurement data is received from the monitored device 500 at a predetermined interval, or may be performed at a timing designated by an administrator or the like.
[0064] In S511, the learning system 400 receives the measurement data from the data management system 300. At this time, the learning system 400 may perform preprocessing on the received measurement data so that the data is in a predetermined format for application to the trained model. The preprocessing may be performed in accordance with the contents of the training data generated in the steps of S412 and S413 in FIG. 4, for example.
[0065] In S512, the learning system 400 uses the first learned model to estimate the state of the monitored device 500 corresponding to the measurement data. Details of this step will be described later with reference to FIG.
[0066] In S513, the learning system 400 generates maintenance information by applying the information about the state estimated in S512 to the second trained model. Details of this step will be described later with reference to FIG.
[0067] In S514, the learning system 400 posts a message including the information on the state generated in S512 and the maintenance information generated in S513 to the SNS system 200. An example of the message will be described later. Note that, in the step of S512, if an abnormality is estimated to have occurred or a sign is estimated as the state of the monitored device 500, the processes of S513 and S514 may be executed. In such a configuration, if an abnormality is estimated to have occurred or a sign is detected, the processes after the step of S513 may be omitted. Alternatively, if an abnormality is estimated to have occurred or a sign is detected, only a post to the SNS system 200 indicating that fact may be made.
[0068] In S515, the learning system 400 generates a control parameter for the monitored device 500 based on the maintenance information generated in S513. The control parameter generated here may be generated corresponding to the state of the monitored device 500, and may be generated based on a table or the like that is predefined corresponding to the generated maintenance information. Alternatively, a stop command for the monitored device 500 may be generated.
[0069] In S516, the learning system 400 transmits a control instruction to the monitored device 500, including the control parameters generated in S515.
[0070] In S521 , the SNS system 200 receives a post from the learning system 400 .
[0071] In S522, the SNS system 200 outputs the post received in S521. The output here may be performed by notifying the management terminal 700 that there is a new post, or may be configured to be output when a specific screen is accessed from the management terminal 700. The content to be output may be the post received in S521 as is, or may be displayed after being converted into a specific format on the SNS system 200 side based on the posted content.
[0072] In S531, the management terminal 700 outputs a post in the SNS system 200. The output here may be a new post on the screen based on a user operation, or may be a notification from the SNS system 200 that can be recognized by the user.
[0073] In S541, the monitored device 500 receives a control instruction from the learning system 400.
[0074] In S542, the monitored device 500 switches the operation content based on the control instruction received in S541.
[0075] Note that steps S515 to S516 in learning system 400 and steps S541 to S542 in monitored device 500 may be omitted depending on the configuration of monitored device 500 and the abnormality that has occurred. In this case, the learning system 400 is configured to simply prompt the device administrator to perform the work without generating control parameters. Also, steps S515 to S516 in learning system 400 may be executed by another device based on a notification from learning system 400.
[0076] (Generating a trained model) 6 is a diagram for explaining generation of a trained model according to the present embodiment. This diagram corresponds to a part of the processing in the training system 400 shown in FIG.
[0077] In this embodiment, two trained models are generated and used: a trained model (first trained model) for estimating the state of the monitored device 500, and a trained model (second trained model) for generating maintenance information for the estimated state.
[0078] In S600, the learning system 400 acquires data. The data acquired here includes measurement data 601, anomaly information 602, and maintenance information 603. As shown in Fig. 5, these data are acquired and managed by the data management system 300, and provided to the learning system 400. Specific examples of each data include the following:
[0079] Measurement data: temperature, vibration, power consumption, continuous operation time, operating points, remaining amount of consumables, control parameters... Abnormality information (string containing "#"): "#abnormality recovery", "#regular maintenance", "#motor overload"... Maintenance information (character strings related to maintenance management): "Motor replacement work was performed", "Periodic inspection was performed"...
[0080] It should be noted that, for the maintenance information, the posted character string may be used as is, or natural language processing such as syntactic analysis and morphological analysis may be performed to extract terms and words for use.
[0081] In S610, the learning system 400 generates learning data to be used in the learning process, using the data acquired in S600. In this embodiment, the learning system 400 generates first learning data 611 to be used when generating a first trained model 631, and second learning data 612 to be used when generating a second trained model 632.
[0082] The first learning data 611 is composed of a pair of measurement data 601 and anomaly information 602. These are associated with each other based on the time stamps associated with each of them. For example, the anomaly information ("#···") is based on an input (post to the SNS system 200) by a maintenance worker. In addition, the anomaly information ("#···") is also based on a post made by the data collection system 100 to the SNS system 200 based on a notification made to the data collection system 100 when the monitored device 500 detects an anomaly. The measurement data 601 for a predetermined period from the time when the anomaly information 602 is input (for example, a range of a certain period before the maintenance work is performed or a certain period before the anomaly is detected) may be associated with the anomaly information 602 and used as the first learning data 611.
[0083] The second learning data 612 is composed of a pair of anomaly information 602 and maintenance information 603. These are included in the same message posted to the SNS system 200, so they may be extracted and associated with each other to form the second learning data 612.
[0084] The first learning data 611 and the second learning data 612 do not necessarily need to be generated at the same time or using the same original data, and may be generated separately. It is preferable that the first learning data 611 and the second learning data 612 are generated for each type and configuration of the monitored device 500. Therefore, for example, the first learning data 611 and the second learning data 612 are generated for a processing device as an example of the monitored device 500.
[0085] In S620, the learning system 400 performs a learning process using the first learning data 611 and the second learning data 612 generated in S610. Note that the first learning process for generating the first trained model 631 and the second learning process for generating the second trained model 632 only need to be performed at the time when a certain amount of learning data is generated, and do not need to be performed synchronously.
[0086] In the first learning process, a learning process is performed using the first learning data 611 and a first learning model 621 based on a predetermined learning algorithm, thereby generating a first trained model 631 that receives measurement data as input and outputs abnormality information. More specifically, when an operating condition, an operation log, or the like is input as measurement data of the monitored device 500, a trained model is generated that outputs abnormality information indicating whether the monitored device 500 is in an abnormal state. In other words, it is a trained model for classifying the measurement data into either an abnormal state or a normal state.
[0087] The classification here may be predefined in response to the anomaly information (character string including "#") posted by the maintenance manager to the SNS system 200. In addition, the learning algorithm related to the classification is not particularly limited, and any learning algorithm may be used.
[0088] In the second learning process, a learning process is performed using the second learning data 612 and a second learning model 622 based on a predetermined learning algorithm, thereby generating a second trained model 632 that receives abnormality information as an input and outputs maintenance information. More specifically, when abnormality information (a character string including "#") is input, a trained model is generated that outputs maintenance information to be performed for the abnormality information. For example, when "#motor malfunction" is input as abnormality information, "replace motor" or "check motor operation" may be output as maintenance information.
[0089] It is also preferable that the first trained model 631 and the second trained model 632 are generated for each type and configuration of the monitored device 500. Therefore, for example, the first trained model 631 and the second trained model 632 are generated for a processing device as an example of the monitored device 500.
[0090] As a result of the above processing, a first trained model 631 and a second trained model 632 are generated and used in processing. As described above, a plurality of first trained models 631 and a plurality of second trained models 632 may be retained depending on the number of times and stages of learning.
[0091] (Condition monitoring using trained models) 7 is a diagram for explaining state monitoring according to this embodiment, which corresponds to a part of the processing in the learning system 400 shown in FIG.
[0092] In S700, the learning system 400 receives the measurement data 701 of the monitored device 500. At this time, the learning system may perform pre-processing to convert the measurement data 701 into a format applicable to the first trained model 631.
[0093] In S710, the learning system 400 performs state monitoring using two models, a trained model (first trained model) for estimating the state of the monitored device 500 described above and a trained model (second trained model) for generating maintenance information for the estimated state. First, measurement data 701 is input to the first trained model 631 to acquire abnormality information for the data. If the abnormal state acquired at this time is classified as indicating that the monitored device 500 has an abnormality or a sign of an abnormality, the abnormality information acquired from the first trained model 631 is further applied to the second trained model 632. As a result, maintenance information corresponding to the estimated abnormal state is output.
[0094] In S720, the learning system 400 uses the acquired anomaly information and maintenance information to post a message to the SNS system 200. Note that the message posted in the posting process in S720 may be the anomaly information and maintenance information acquired in S710 posted as is, or may be posted after undergoing a predetermined conversion.
[0095] For example, for maintenance information, the character string of the maintenance information acquired in S720 may be embedded in a predetermined sentence pattern so that the user (e.g., the device manager) can easily understand it. For example, when "motor replacement" is acquired as maintenance information, this may be applied to a predefined sentence pattern "Please do XX" to generate a message "Please replace the motor." The sentence pattern may be predefined in accordance with the above classifications, etc.
[0096] Alternatively, the second trained model 632 itself or a trained model separately provided thereafter may be configured to generate maintenance information as a message to be posted. For example, when the second trained model 632 itself generates a message to be posted, when "#motor malfunction" is input as abnormality information, the model may be configured to output "Please check the motor operation" or "Prepare for motor replacement" as maintenance information. As a learning algorithm for handling such natural language, a convolutional neural network (CNN) using a framework such as Tensorflow is known, and these can be applied.
[0097] With the above-mentioned configuration, the status monitoring system 1 can monitor the status of the monitored device 500 based on the measurement data from the monitored device 500. When an abnormality or the like is detected, the system can notify the device manager of the abnormality and the operation required according to the detected abnormality, and prompt the device manager to perform the operation. The device manager can grasp the change in the status without individually checking and analyzing various measurement data generated by the operation of the monitored device 500. As a result, the operation management of the monitored device 500 becomes easier, the workload is reduced, and the work to be performed when an abnormality occurs can be easily grasped. In addition, not only abnormalities that have already occurred but also signs of abnormality can be captured, so that a response or preparation can be made before a failure occurs, and further, the occurrence of downtime due to a breakdown or the like can be reduced.
[0098] <Other embodiments> In addition, in the present invention, a program or application for realizing the functions of one or more of the above-mentioned embodiments can be supplied to a system or device via a network or a storage medium, etc., and one or more processors in a computer of the system or device can read and execute the program, thereby making it possible to realize the present invention.
[0099] Moreover, it may be realized by a circuit that realizes one or more functions (for example, an application specific integrated circuit (ASIC) or a field programmable gate array (FPGA)).
[0100] In addition, in this specification, the terms "first" and "second" are used merely to distinguish from other elements, and are not intended to be interpreted as being limited to a specific element. Therefore, these terms may be interpreted appropriately depending on the configuration to which the present invention is applied.
[0101] As such, the present invention is not limited to the above-described embodiments, and the present invention also contemplates mutual combinations of the various components of the embodiments, as well as modifications and applications by those skilled in the art based on the descriptions in the specification and well-known technologies, and these are included in the scope of protection sought.
[0102] As described above, the present specification discloses the following: (1) A status monitoring system (e.g., 1) for a monitored device (e.g., 500), A first system (e.g., 200) for providing a social networking service; a second system (e.g., 400) for determining a state of the monitored device using measurement data (e.g., 701) during operation of the monitored device; The present invention relates to a method for manufacturing a computer-implemented ... The first system comprises: A means (e.g., 11, 15, 16) for accepting a posting of a first message including first character string information (e.g., 602) indicating a state occurring in the monitored device and second character string information (e.g., 603) indicating the content of the maintenance performed in response to the state; A means (e.g., 11, 15, 16) for providing posted messages to other systems (e.g., 300, 400); A means (e.g., 11, 14) for outputting a message posted on the social networking service; having The second system comprises: A means (e.g., 11, 15, 16) for acquiring measurement data during operation of the monitored device; A means (e.g., 11) for acquiring first character string information (e.g., 721) and second character string information (e.g., 721) corresponding to new measurement data (e.g., 701) using a trained model (e.g., 631, 632) that receives measurement data as input and outputs first character string information and second character string information corresponding to the measurement data; A means (e.g., 11, 15, 16) for posting a second message (e.g., 721) including the acquired first string information and the second string information to the first system; having A condition monitoring system, wherein the trained model is generated by learning using training data composed of first string information and second string information contained in a message acquired from the first system and the measurement data. According to this configuration, the status monitoring system 1 can monitor the status of the monitored device 500 based on the measurement data from the monitored device 500. When an abnormality or the like is detected, the system can notify the device manager of the abnormality and the operation required according to the detected abnormality, and prompt the device manager to perform the operation. The device manager can grasp the change in the status without individually checking and analyzing various measurement data generated by the operation of the monitored device 500. As a result, the operation management of the monitored device 500 becomes easier, the workload is reduced, and the work to be performed when an abnormality occurs can be easily grasped. In addition, not only abnormalities that have already occurred, but also signs of abnormality can be captured, which allows for response and preparation before a failure occurs, and further allows for reduction in downtime caused by breakdowns and the like.
[0103] (2) The condition monitoring system includes a third system (e.g., 100) that collects measurement data during operation of the monitored device, The third system comprises: A means (e.g., 11, 15, 16) for providing measurement data collected from the monitored device to another system (e.g., 300); a means (e.g., 11, 15, 16) for posting a third message including first character string information indicating the content of a notification when a notification indicating the occurrence of an abnormality is received from the monitored device to the first system; The condition monitoring system according to (1) above, According to this configuration, when the data collection system 100 acquires information regarding the detection of an anomaly, it is possible to notify the device administrator by posting the information to the SNS system 200. Furthermore, the information regarding the posting can also be used to generate learning data used in the learning process in the learning system 400.
[0104] (3) The status monitoring system according to (1) or (2), wherein the first character string information is a character string to which a hashtag has been added. This configuration makes it possible to identify and extract abnormal conditions from messages posted on SNS, and improves the convenience of information extraction when using learning data and classification of abnormalities.
[0105] (4) The second system, A means (e.g., 11) for generating the learning data by associating a time stamp given to a message acquired from the first system with time information of a measurement time indicated in the measurement data; A means (e.g., 11) for generating a trained model (e.g., 631, 632) that receives measurement data as input and outputs first character string information and second character string information corresponding to the measurement data by performing a training process using the training data; The condition monitoring system according to any one of (1) to (3), further comprising: According to this configuration, the collected measurement data, abnormality information, and maintenance information can be collected as appropriate and used in the learning process, thereby updating the learned model and enabling accurate condition monitoring tailored to the monitored device.
[0106] (5) The second system, A means (e.g., 11) for generating a control parameter for the monitoring target device based on the acquired first character string information and the acquired second character string information; A means (e.g., 11) for controlling the operation of the monitored device using the generated control parameters; The condition monitoring system according to any one of (1) to (4), further comprising: According to this configuration, it becomes possible to control the operation of the monitored object in response to the detected abnormality or symptom.
[0107] (6) a first system (e.g., 200) providing a social networking service; a second system (e.g., 400) for determining a state of a monitored device (e.g., 500) using measurement data (e.g., 701) during operation of the monitored device; A status monitoring method for a status monitoring system (e.g., 1) comprising: In the first system, A step (e.g., S333) of accepting a posting of a first message including first character string information (e.g., 602) indicating a state occurring in the monitored device and second character string information (e.g., 603) indicating the content of the maintenance performed in response to the state; providing the posted message to another system (e.g., S332, S334); A step of outputting the posted message on the social networking service (e.g., S522); having In the second system, A step of acquiring measurement data during operation of the monitored device (e.g., S511); A step (e.g., S512, S513) of acquiring first character string information and second character string information corresponding to new measurement data using a trained model (e.g., 631, 632) that receives measurement data (e.g., 701) as input and outputs first character string information (e.g., 721) and second character string information (e.g., 721) corresponding to the measurement data; A step of posting a second message including the acquired first string information and the second string information to the first system (S514); having A condition monitoring method, wherein the trained model is generated by learning using training data composed of first string information and second string information contained in a message acquired from the first system and the measurement data. According to this configuration, the status monitoring system 1 can monitor the status of the monitored device 500 based on the measurement data from the monitored device 500. When an abnormality or the like is detected, the system can notify the device manager of the abnormality and the operation required according to the detected abnormality, and prompt the device manager to perform the operation. The device manager can grasp the change in the status without individually checking and analyzing various measurement data generated by the operation of the monitored device 500. As a result, the operation management of the monitored device 500 becomes easier, the workload is reduced, and the work to be performed when an abnormality occurs can be easily grasped. In addition, not only abnormalities that have already occurred, but also signs of abnormality can be captured, which allows for response and preparation before a failure occurs, and further allows for reduction in downtime caused by breakdowns and the like.
[0108] Although various embodiments have been described above with reference to the drawings, it goes without saying that the present invention is not limited to such examples. It is clear that a person skilled in the art can come up with various modified or revised examples within the scope of the claims, and it is understood that these also naturally belong to the technical scope of the present invention. Furthermore, the components in the above-mentioned embodiments may be arbitrarily combined within the scope of the invention. [Explanation of symbols]
[0109] 1. Condition monitoring system 10...Information processing device 11...Control section 12...Storage section 13...Operation unit 14...Display section 15…External IF 16…Communications Department 100...Data Acquisition System 200…SNS system 300...Data management system 400…Learning System 500: Monitored device 600…Maintenance terminal 700: Management terminal
Claims
1. A status monitoring system for a monitored device, comprising: A first system for providing a social networking service; a second system for determining a state of the monitored device using measurement data during operation of the monitored device; The present invention relates to a method for manufacturing a computer-implemented ... The first system comprises: a means for accepting a posting of a first message including first character string information indicating a state occurring in the monitored device and second character string information indicating a maintenance operation performed in response to the state; a means for providing posted messages to other systems; A means for outputting a message posted on the social networking service; having The second system comprises: means for acquiring measurement data during operation of the monitored device; A means for acquiring first character string information and second character string information corresponding to new measurement data using a trained model that receives measurement data as input and outputs first character string information and second character string information corresponding to the measurement data; means for posting a second message to the first system, the second message including the acquired first string information and the second string information; having A condition monitoring system, wherein the trained model is generated by learning using training data composed of first string information and second string information contained in a message acquired from the first system and the measurement data.
2. the condition monitoring system is configured to include a third system that collects measurement data during operation of the monitored device, The third system comprises: A means for providing the measurement data collected from the monitored device to another system; a means for posting a third message to the first system when a notification indicating the occurrence of an abnormality is received from the monitored device; and The condition monitoring system of claim 1 .
3. The status monitoring system according to claim 1 , wherein the first character string information is a character string to which a hashtag is added.
4. The second system comprises: a means for generating the learning data by associating a time stamp given to a message acquired from the first system with time information of a measurement time indicated in the measurement data; A means for performing a learning process using the learning data to generate a learned model in which measurement data is input and first character string information and second character string information corresponding to the measurement data are output; The condition monitoring system of claim 1 further comprising:
5. The second system comprises: means for generating a control parameter for the monitoring target device based on the acquired first character string information and the acquired second character string information; a means for controlling an operation of the monitored device using the generated control parameters; The condition monitoring system of claim 1 further comprising:
6. A first system for providing a social networking service; a second system for determining a state of the monitored device using measurement data during operation of the monitored device; A status monitoring method for a status monitoring system comprising: In the first system, receiving a posting of a first message including first character string information indicating a state occurring in the monitored device and second character string information indicating a content of maintenance performed in response to the state; providing the posted messages to another system; outputting a posted message on the social networking service; having In the second system, acquiring measurement data during operation of the monitored device; A step of acquiring first character string information and second character string information corresponding to new measurement data using a trained model that receives measurement data as input and outputs first character string information and second character string information corresponding to the measurement data; posting a second message to the first system, the second message including the acquired first string information and the second string information; having A condition monitoring method, wherein the trained model is generated by learning using training data composed of first string information and second string information contained in a message acquired from the first system and the measurement data.
Citation Information
Patent Citations
Data transfer processor, device equipment, network system, data transfer method and storage medium
JP2002215484A
Image formation device, print execution application program, and printing system
JP2014104625A
